Reinforcement Learning and World Model for Autonomous Simulation Intern - 2027
Salary not listed • China, Shanghai • Internship
Posted 8d ago
Remote Work Policy
On-site
Employment Type
Internship
Categories
AI Research Engineer
About the job
We are seeking a hardworking intern with expertise in Reinforcement Learning and Multi-Modal World Simulation Models to advance ML-centric autonomous driving and Physical AI solutions. The role focuses on model-centric RL, understanding world simulation models, and applying state-of-the-art algorithms to real-world applications, enabling vehicles to interpret, anticipate, and respond effectively in dynamic environments. This is an opportunity to contribute to the future of intelligent driving, where innovation drives real-world impact. If you are passionate about building simulation technologies and systems that learn, adapt, and think, we encourage you to apply. Join a team where your contributions are vital to accelerating autonomous vehicle growth with advanced solutions.
Responsibilities
- Develop and refine multi-modal world models and integrate them into the simulation system.
- Train and evaluate self-supervised latent dynamics and sensor generation models for trajectory prediction, goal-conditioned ego control, and sensor data synthesis.
- Explore and prototype hybrid architectures combining world models, generative models (e.g., diffusion, flow matching), and policy gradients for realistic simulation.
- Collaborate with End-to-End Driving Model teams to deploy world-model-based policies to simulated RL environments.
- Contribute to system development for continuous learning and simulation adaptation (Sim2Real transfer).
Requirements
- Pursuing PhD in Computer Science, Machine Learning, or a related field, with a background in neural rendering, robotics, or simulation.
- Strong understanding of reinforcement learning (policy gradients, actor-critic, offline RL).
- Familiarity with visual representation learning and 4D scene representation (NeRF, Gaussian Splatting, occupancy networks, contrastive, masked modeling, or generative world simulation) for world simulation.
- Experience building large-scale training pipelines with temporal consistency and simulation data replay.
- Publications or open-source contributions in RL, model-based control, or autonomous systems.
- Passion for developing learning systems that can 'imagine' and plan in the real world.